Lithium-ion batteries are key components for electric mobility and renewable energy systems. Their effective use in such applications requires precise tools for real-time diagnosis and refined energy management techniques. While battery characterization and modeling are closely related, they are often tackled in isolation due to the challenge of developing models that encapsulate complex physicochemical processes. In response, this paper presents a new approach for an on-line, non-invasive characterization technique of the battery impedance based on power steps, which are a typical process in regular battery operation. The proposed impedance representation is linked to the underlying physicochemical phenomena that drive the operation of a battery, which offers a real-time diagnosis tool to infer insights into underlying phenomena such as lithium anode depletion and phase transitions within both electrodes. This characterization proves instrumental in identifying the key parameters for an equivalent-circuit model. The proposed methodology is proven to be resilient and accurate, given the comprehensive dataset of 1600 experiments used to fit the impedance parameters. Finally, we scale up the cell-level model to represent a commercial 38.4 kWh battery pack. The results prove the robustness of the model, providing a normalized root mean squared error lower than 0.5% across all steady-state and dynamic validations. In essence, this paper proposes a methodology that combines real-time access to underlying battery dynamics with accurate performance simulation.
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号